It is obvious and natural to identify a person based on his/her biological trait and we do it since our birth. But now,we came to know how our brain does this activity. Iris has always been one of the most precious, trusted and accurate meansof recognition due to its unique and ageless features. Traditional approach of iris recognition system had to face a lot ofproblems in image capturing, pre-processing and extracting features. But now, Artificial Intelligence is the new boss and it has brought a revolution by introducing Convolutional Neural network, which possesses a resemblance to the human brain.This paper has made use of a CNN model namely Alex Net, for feature extraction, which automatically extracts a huge matrix of features. These features were passed in SVM (Support Vector Machine) classification to classify images among the correct classes. The proposed method preferred to use iris database obtained from IITD to conduct the research. Here, CASIA-Iris-Interval V4 and the UBIRIS.v2 datasets are used for the experimentation. With the Alex-Net, the possible accuracy rates on the UBIRIS.v2 and CASIA-Iris-Interval V4 datasets are 92.76%, and 97.05% respectively.